16 _Potrolaum Retinery Distillation balance, will reduce the (5-95) Gap which isthe separation. ‘AC this. point, the similarity between discrete-component distillation and erude petcoleum fractionation should be Tess unclear. ‘As more operating data has become available, Packie's work is now generally considered to be on the conservative side? that is, his procedure usually predicts a smaller gap than is actually attained in practice, The author believes that the degree of conservatism inherent in Packie is not ‘excessive and recommends Its use for design work. ft is almost always trae thet erude distillation units are required to perform different operations than that for which they were designed, For this reason, the design “Tat” contained in Packie becomes quite useful to the owner in later years. Atmospheric Tower Charge Data The feed to the stmospheric tower is the enude oll to be processed into the various products required. Before the design material balances can be developed, it will Mrst be necessary t0 derive some basic pliysical property data for the crude. From this, one can then estimate the total distil late production and product distribution in the atmospheric tower. In definitive process design work andfor in refinery operations analysis, the engineer usually has access to complete erude assay. Indeed, he should not undertake & definitive design without one. On the other hand, feasibility studies and/or order-of magnitude work is often performed with little more than a whole crude gravity and TBP curve In either case, the first step isto characterize the crude in ‘order to facilitate leter calculations. The minimum informa- tion required is (a) wholecrude TEP curve, (b) whole-crude API gravity and (¢) whole-crude light ends analysis. ‘Additional information which is highly desirable is an API gravity and molecular weight study of narrow cuts of| the whole crude. From this can be derived a plot of volume percent versus weight percent and mid-volume percent ver sas molecular weight. If this data is not available experi mentally, it can be calculated. ‘An experimental determination of the FV curve, at least at atmospheric pressure and preferably at one or two higher pressures covering the anticipated range of opera tions is also desirable, Maxwell (4), Nelson and Edmister have all published procedures for converting a whole crude TBP curve to an atmospheric EFV curve and for estimating the EFV temperatures at pressures above atmospheric. In the petroleum industry, there is a good deal of argument about how to estimate EFV behavior of erudes since the total vaporization in the tower is the sum of the equil ‘rium flash vapor plus some stripout from the equilibrium flash liquid, Stripout versus stripping steam correlations are stripout by making a heat balance around the bottoms stripping section, bus this is not possible in a design situa tion. For these reasons, the author strongly recommends obtaining experimental data, particularly when working with unfamiliar ceudes. The cost of obtaining accurate equilibrium data is trivial when competed to the possible penalties involved in incorrectly estimating yields or mis sizing equipment. However, if one is limited to the minimum data, the required information can be derived by using the following procedures for erude oil characterization. Calculation Procedure for Characterizing Crude Oils Giver: Whole-crude atmospheric TBP curve Whole-crude gravity Front ond analysis for light ends Calculate: Volume percent versus weight percent Mid-volume precent versus molecular weight 1. Calculate the characterization factor, K, for the whole crude using the techniques of Maxwell, Section 2 a, Calculate the volume average boiling point (VABP) using the 20, 50 and 80 volume percent TBP tem> peratures. bb. Calculate the 10 to 70 slope of the wholecrude ‘TBP curve, cc. Using the proper correction factor, convert VABP to mean ABP. 4d. K is found as a function of mean ABP and API gravity by use of Winns’ nomogram (5). 2. Assuming that K remains constant, calculate the mol cecular weight and API gravity for various boiling-range cuts. a. For the TBP range (initial boiling point to 200 degrees F), calculate mean ABP, API gravity and ‘molecular weight. Based on 100 barrels of whole ‘rude, calculate and tabulate: barrels vapor pounds vapor moles vapor molecular weight of vapor : b. For the THP range (200 to 300 degrees F), repeat Step 2a . For the TBP range (300 to 400 degrees F), repeat ‘Step 22, and so on until the entire cnide range has been covered. OR ROMER EE OOOO LL eecanaae ¢ ¢ ¢ yawe we ’ ’ o ° » euwouvdew 4, From these calculations, draw the following 1, Volume percent over versus weight percent 2. Mid-volume percent versus molecular weight of vapor. 2, Caleuate the atmospheric pressure EFV eune for thé whol crude using Packie procedue, Plt this cure on the sme chart athe whole crude TAP cure Estimate of Material Balance Asis the casein any proces:-design problem, the defn tion of the material balance i the fist and most important step. In this portion of the diseusion, vacious methods for estimating product yields from the crude petroleum feed will be explored “The design material alance is determined by the prod uct characteristics required by the owner and by the mount of crude vaporization which wil occur at the con- ditions of temperature and pressure existing in the Mash zone, Since most crude distillation units conta both atmos- pherie and vacuum towers, economic considerations usually favor maximizing distillate yield from the atmospheric sec tion in order to minimize the load on the vacuum section. This may not always be tru in an operating situation where the vacuum tower might have unused capacity when the atmospheric tower is operated for maximum distillate pro- duction. In that case, one would balance atmospheric and vacuum distillate progvtion in such a way as to maximize crude throughput. In designing atmospheric towers where the bottoms liquid is sold directly a8 fuel oil, the most economical approach in this special case sto vaporize only the required distillate products and to allow the remaining potential distillate 1 be yielded, unvaporized, with the bot foms. This situation arises often ia foreign refineries, usually in the producing areas, which run erude primarily for iocal fuels requirements and for bunkering tankers. Thus, there is just no economic driving force for yielding anything heavier than. diesel fuel as a distin In practice, the question of optimizing the relative dis- tillate yield between the atmospheric and vacuum towers will be settled on an economic basis and must be resolved ptior to commencing definitive design work. This type of analysis will depend upon econonc factors within the pat- ticular company and/or plant site under study and is out side the scope of this work ‘Total Disilate Yield ‘The total distillate yield is found by calculating the ‘vaporization of the crude which will occur at the conditions Atmospheric Tower 17 TABLE 24 Tray Requirements in Atmospheric Crude Towers Separetion Number of Trays] Light naphtha to heavy naphtha 6108 Heavy naphtha to light distillate 6108 Light distillate to heavy distilate 4106 Heavy distillate to atmospheric ges oll | 4 106 Flash zone to first draw tray 3104 ‘Steam and reboiled stripping sections 4 ‘of temperature and hydrocarbon partial pressure existing in the flash zone. The principles of this analysis are discussed as follows, Flash Zone Pressure Flash 2one pressure should be set at the minimum pos- sible level. This will maximize the erude vaporization or, if fess Than maximum vaporization is desired, will minimize ‘the operating temperature level. ‘The datum pressure in atmospheric tower design is the lowest permissible pressure in the towers’ reflux accumula: tor. Since, in many cases, the tower overhead product stream cannot be totally condensed at a reasonably low pressure, there will often be a vapor distillate which must bbe removed from the tower system and sent either to the plant fuel gas system or to fusther processing. This is nor mally done by a compressor located as near as possible to ‘he tower. In most refineries, tis distillate vapor is fed toa 2s plant for recovery of propane and heavier. In refineries running heavy erudes, the volume of recoverable light ends may be too small to justify recovery facilities. In this case, the distillate vapor is sent to the plant fuel gas system. The author has observed one opéiation where the distillate vapor was flared. This required a considerably higher reflux drum pressure than if compressors were used to feed other systems, Admittedly, this was a rare operation. Fora grass roots design a reflux drum pressure as low as 0.2 psig is practical, provided that suetion lines to the compressor can bbe kept short. If the installation involves tying a new tower into an existing compressor which may be located a consid- rable distance away, a higher pressure may be required. ‘The nogmal cange of reflux drum pressures used is 0.2 10 1.0 psig ‘The design pressure drop across the condenser(s) and associatéd piping should be set at 5 to 7 psi. When the system is clean, the normal pressure drop will be on the18 Petroleum Refinery Distillation EXTERNAL HEAT QUANTITIES LEAVING FLASH, ZONE Qlirz= (QF +Qsw), BTU/ hr. Q'ore= Qu, BTU/ by aarr =QoIF2, BTU/hr F Tro Ten On Pro. Pra PREHEATED Yo Veni FEED FROM i Ue EXCHANGE 7 i eet Sw. STRIPPING STEAM ay ED=TOTAL DISTILLATE PRODUCTS EXCLUSIVE Lo= OVERFLASH Vso= STRIPOUT HEAT INPUT TO PROCESS Qro=Qi=(QvtQilro Qrr1=Q2=(Qv+Qu)H2 rm Qi=Q2, BTU/hr. HYDROCARBON PARTIAL PRESSURE IN FLASH prc |[vacen=(v'+vs0)] /(vaces +Sw)} Pez, PSA WHERE ALL QUANTITIES ARE MOLES. Lo) Sw Lor (Cw=Tra7 30°F ‘BOTTOMS Ow OF OVERFLASH ZONE Figuee 2.11. Flesh zone heat and material balance, order of 3 psi. However, as the exchangers become fouled. the extra drop allowed will be available. Most atmospheric towers have 25 to 35 trays between the flash zone and the tower top. Recommended ranges for the number of trays in various sections of the tower are given in Table 2.1. Using these numbers of trays and reason able levels of heat recovery will provide adequate fractiona- tion between products. The allowable pressure drop for trays will be in the range of 0.1 to 0.2 psi per tray. In the design example, 2 pressure drop of S psi was taken between the fash zone and the tower top. {At this point, the flash zone pressure is set as the sum of the reflux drum pressure and the combined pressure drops across the condensers and trays above the flash zone. A pressure drop of 5 psi between the flash zone end the furnace outlet is recommended. « « aaa < t OM ere ¢ ¢ ¢